Predictive Analytics Model for Small and Medium Enterprises in Kenya, Forecasting on Supply and Demand

Abstract

Predicting business operations is a critical task for small and medium enterprises (SMEs). With increased unpredictability in the business environment, small enterprises find themselves in the receiving end simply because they do not have the tools in decision-making like their counterparts who have established business decision-making tools. With increased use of ICT, SMEs can now tap into the power of data to support decision-making. Transactional data like sales, purchases, payments, service requests, invoices, purchase orders, delivery notes, repairs, and maintenance notes are collected by SMEs and are readily available. Currently, SMEs in Kenya have limited on non-existent quantitative forecasting systems in place, predictive analytics is a preserve of well-funded, large, well-established or international companies. SME owners can now benefit from the power of predictive analytics in areas like sales, purchases, business lead generation, recommendations systems, and risk prediction. With predictive analytics, SME owners can have added confidence in decision-making to help propel their business to successful ventures. Predictive analytics algorithms like clustering algorithms, linear regression, and classification algorithms can be used to aid SME owners and managers gain insights into their business by identifying relationships and associations between the various variables in their business or identification of trends. SMEs in Kenya, have in the past run out of business due to wrong decisions associated with lack of information regarding their customers, their business, or the business sector as a general. SMEs' contribution to Kenya’s GDP growth is vital and the use of technology and ICT could mitigate the challenge of access to information that SMEs have. The use of technology to predict business operations and performance is the next frontier in ensuring business sustainability, job security, and a good business environment. This research aimed to solve this information gap by designing a demand and supply forecasting model that equips SME owners and managers with insightful information about the demand and supply of the products they are selling, helping them make informed decisions based on their sales and purchase data. The study was carried out on five (5) SMEs in Nairobi and Muranga who had access to ICT infrastructure and had some knowledge of digital book-keeping methods but had no forecasting or prediction systems in place. A fully functional web passed demand and supply forecasting model accessible via a browser was designed. A beta test was conducted by the five respondents with positive feedback. The majority of the respondents, as the study found out, derived great value from forecasting their demand and supply and were able to stock right and meet their clients’ needs better, the forecasting model was developed using the agile prototyping method of software development with cascading style sheets (CSS), hyper-text markup language (HTML), react graphical user interface and an R based forecasting engine. The designed system allowed the users to interact with a user-friendly graphical user interface on either a mobile device or a computer allowing more freedom and flexibility in accessing the platform. It allowed SMEs to upload their sales and purchase data in a predefined format that the forecasting model consumes and provide accurate demand and supply forecasts and provided forecasting accuracies in the 95% quartiles for all the SMEs sampled. The model was evaluated by the same SMEs with 100% of them indicating that they would be relying on the forecasting model for all future forecasts. The designed model allows SMEs to benefit from the demand and supply forecasts irrespective of its sector and the type of enterprise engages in.